胸腔积液肿瘤标志物的PCA和OPLS-DA模型对肺癌的诊断价值研究

目的探讨联合检测胸腔积液中癌胚抗原(CEA)、细胞角蛋白19片段(Cyfra21-1)、神经元特异性烯醇化酶(NSE)、鳞状细胞癌抗原(SCC)和胃泌素释放肽前体(Pro-GRP)对肺癌的诊断价值。方法选择西南医科大学附属医院2014年2月至2016年3月127例肺癌患者为肺癌组,随机选择154例良性胸腔积液患者为对照组。采用电化学发光法和化学发光法测定胸腔积液中CEA、Cyfra21-1、NSE和SCC和Pro-GRP水平。应用主成分分析(PCA)法和正交偏最小二乘判别分析(OPLS-DA)法评估联合检测胸腔积液5种肿瘤标志对肺癌的诊断价值。结果肺癌患者胸腔积液中CEA、Cyfra21-1、NSE,SCC和Pro-GRP水平均显著高于对照组,差异有统计学意义(P0.01)。PCA模型表明,两组个体的代谢轮廓各不相同,有分离趋势。OPLS-DA能较好地鉴别肺癌和肺部良性疾病,具有88.6%(124/140)的灵敏度、90.9%(140/154)的特异性和89.8%(264/294)的准确率。结论基于胸腔积液CEA、Cyfra21-1、NSE和SCC和Pro-GRP的PCA和OPLS-DA模型有助于肺癌的鉴别诊断。

Diagnostic value of tumor markers in pleural effusion for patients with lung cancer based on PCA and OPLS-DA models

Objective To evaluate the combined diagnostic values of CEA ,Cyfra21‐1 ,NSE ,SCC and Pro‐GRP for patients with lung cancer .Methods Totally 127 cases with lung cancer were selected in the hospital from Feb .2014 to Mar .2016 ,which were di‐vided as lung cancer group .And 154 patients with benign pleural effusion were selected as control group .Concentrations of CEA , Cyfra21‐1 ,NSE ,SCC and Pro‐GRP in pleural effusion were measured by the electrochemiluminescence and chemiluminescence as‐says .The methods of principal component analysis (PCA) and orthogonal partial least squares‐discriminant analysis (OPLS‐DA) were used for evaluation the diagnostic performances of five tumor markers in pleural effusion for patients with lung cancer .Results Concentrations of CEA ,Cyfra21‐1 ,NSE ,SCC and Pro‐GRP in pleural effusion in the lung cancer group were significantly higher than that in the control group (P< 0 .01) .PCA model revealed that there was different metabolic profiling between lung cancer group and control group ,showing a tendency of discrimination .The lung cancer and other benign lung diseases were effectively dis‐criminated by the OPLS‐DA model with 88 .6% (124/140) of sensitivity ,90 .9% (140/154) of specificity and 89 .8% (264/294) of accuracy ,respectively .Conclusion Based on tumor markers of CEA ,Cyfra21‐1 ,NSE ,SCC and Pro‐GRP in pleural effusion ,the PCA and OPLS‐DA models are helpful for the diagnosis of lung cancer .